Summary Current Population Survey (CPS) data document the persistence of food insecurity among America's children despite an array of government programs geared toward improving food access and nutrition. The significance of child food insecurity (or CFI) lies in its negative developmental, health, educational, and later-life consequences. Yet the members of the US population arguably most vulnerable to such consequences?homeless children and their families?are not captured by the CPS study design. Hence, much of what is known about homelessness and food insecurity comes from research that is local in scope, based on shelter or clinical samples, or restricted to certain kinds of homeless youth (e.g., adolescent runaways). Differences across studies in methodology and substantive emphasis, coupled with an atheoretical bent, further fragment this research and limit its integration into the larger body of food insecurity scholarship. To remedy these problems, we propose a comprehensive analysis of food insecurity among children in homeless and precariously housed families. Guiding our research is a theoretical model of family resource allocation in which risk and protective factors, operating partially through parents' managerial capacity, influence their offspring's access to food. Older (1996) but unexploited data from the National Survey of Homeless Assistance Providers and Clients (NSHAPC) are used to evaluate the model. The NSHAPC client interview contains selected child and adult measures drawn from the CPS food security module that allow us to (1) estimate the extent of CFI among homeless families, including comparisons to all US households with children, the subset of those households below the poverty line, and precariously housed families in the NSHAPC sample; (2) identify which theoretically relevant variables predict the likelihood of homeless and precariously housed families having food-insecure children, with attention to the relative importance of parental capital, parental vulnerabilities, family composition, and instrumental behaviors by parents (to achieve housing stability, obtain government assistance, and facilitate their children's institutional integration); (3) account for the food acquisition strategies of the sample families, detailing the number and types of food sources employed and the degree to which different strategies are associated with the factors noted in aim 2; and (4) probe the relationship between parent food insecurity and CFI within homeless and precariously housed families, which may be concordant (when parent and child food security statuses align) or discordant (when they do not). We address these four aims by examining the experiences of the 738 parent respondents (487 currently homeless, 251 precariously housed) in the NSHAPC sample and the 1,614 children under age 18 in their care. For the initial aim, 1996 CPS food security microdata will enable the necessary comparisons. Most of the descriptive and multivariate analyses, however, will examine variation among the NSHAPC families (homeless vs. precariously housed, food-insecure vs. secure).